I see this completely as Apple stoping competition vs any type of safety. We already get prompted a million times by TCC for "node" and other nonsense you cannot even understand or reason about. If they cared about safety they would provide a dedicated sandbox environment or other ways to better introspect what agents are doing as they run without stopping them.

It’s already difficult enough, unless Apple removes the option completely I don’t see how this improves anything. Users don’t read prompts.

I've heard too many stories of personal friends getting into trouble from just recklessly running agents on their Macs (giving Full Disk Access perms) for me to balk at this. Those of us who are technical enough can find ways around, but the public needs to be protected from themselves at this moment. I can't believe I'm actually saying that, but here we are. I blame the AI companies.

I already struggle with terminal and scripts with full disk access, this will probably make it worse.

I suspect this reactive nonsense, is part of apple's plan to ios-ify macos.

So, some people are suffering the consequences of their free descision to run known dangerously unreliable programs, and everyone sensible enough to not run them has to pay the price.

A.k.a. this is why we can't have nice things.

[flagged]